HIIT seminars in spring 2007 will be held in hall **B222** of Exactum,
on Fridays starting at 10:15 a.m. Coffee available from 10.
Fri Apr 20
Kai Puolamäki
Using eye movements in information retrieval
Abstract:
A traditional information retrieval (IR) system tries to present a
user material that she would find most relevant, based on a typed
query. It would be useful if the system would be able to infer from
the behaviour of the user whether she finds the documents presented by
the IR system relevant, the ultimate goal being to gain information of
the relevance of yet unseen documents.
We have approached the problem by measuring and analyzing the eye movements of users during IR tasks in controlled experiments.
We have constructed models based on discriminative learning which can infer whether the user finds the text she is reading relevant, based on eye movements of the user and other sources of relevance feedback. Furthermore, we have constructed a system where IR query is inferred from eye movements measured when the user is reading text during an IR task.
This ongoing work has been done done jointly with Samuel Kaski, Jarkko Salojärvi and Antti Ajanki from 2004 onwards.
Last updated on 19 Apr 2007 by Teija Kujala - Page created on 20 Apr 2007 by Teija Kujala